Hybrid Temporal-SNR Multiple Description Coding for Peer-to-Peer Television

In this paper we present Hybrid Temporal/SNR Scalable Multiple Description Coding. For efficient Peer-toPeer streaming more than two descriptions are required. In our P2P-TV video streaming network, we also need scalability and error resilience. To increase the number of descriptions we use SNR enhancement layers with channel codes (MD-FEC). The robustness against failures in the P2P network, such as packet losses and peer-failure, is increased by either employing odd/even frame MDC or by temporal layering. We show different hybrid MDC schemes and their respective rate distortion curves and redundancy figures, and discuss the impact of packet-loss and peer-failures on these schemes. Given a peer-to-peer network with certain dimensions, loss rates and failure rates and scalability demand we then optimize the quality of the hybrid MDC streaming over a P2P-TV network.